Problem Solver • Building on-device AI & high-scale distributed systems
I am an AI & Data Science Pre-final Year student at Vishwakarma Institute of Technology, Pune (CGPA: 8.6). I focus on engineering reliable, low-latency software solutions where modern AI models meet distributed backend infrastructure.
My engineering work specializes in on-device local AI solutions (quantized SLMs with Ollama and Hugging Face), RAG (Retrieval-Augmented Generation) pipelines with vector search, and high-throughput event streaming systems using Apache Kafka and Redis.
I am deeply committed to problem solving with 400+ DSA problems solved across platforms, a 7× Hackathon Finalist, and the winner of Techathon 3.0 for developing an offline AI tutor. Outside of competitive programming, I serve as the Cloud Coordinator at MLSC VIT Pune, leading hands-on cloud and developer workshops.
Leading cloud initiatives, technical workshops, and hands-on bootcamps covering cloud architecture, Azure, and AI deployments for students and community members.
Built and maintained core web portals for campus student services, ensuring uptime, responsive UI, and smooth digital workflows for institutional events.
Engineered interactive frontend user interfaces, optimized rendering pipelines, and integrated RESTful backend APIs.
Curated technical articles, managed communication pipelines, and coordinated editorial outreach.
Python
C
C++
JavaScript
TypeScript
SQL
Bash
Ollama
Hugging Face
PyTorch
FastAPI
Flask
FastAPI
React
Node.js
Express
HTML5 & CSS3
Apache Kafka
Redis
MySQL
MongoDB
Docker
Git
GitHub
GitHub Actions
AWS
Linux
Postman
GitHub Foundations ↗
AWS Cloud Practitioner
5-Day AI Agents (Google)
SQL Fundamentals (DataCamp)
An AI-powered interview and candidate screening platform built for high-volume recruitment. Automates campus hiring workflows through structured realtime voice interviews with Gemini Live, intelligent resume parsing, multi-agent evaluation scoring, and explainable rubric assessments.
An on-device offline AI tutor designed specifically for educational equity in remote environments without dependable internet connectivity. Powered by Gemma 3 & Qwen 3 via local quantized model inference, delivering real-time multimodal tutoring, interactive quizzes, and personalized concept reinforcement.
High-performance Retrieval-Augmented Generation (RAG) platform capable of answering complex user inquiries across dense, multi-hundred-page PDF documentation. Implements FAISS similarity search, chunk optimization, and LLaMA 3.1 with a FastAPI backend containerized with Docker.
Event-driven transaction processing system designed to handle continuous streams of concurrent banking and payment ledger events. Utilizes Apache Kafka for durable distributed message queuing and Redis for in-memory caching and state verification.